As CAR T-cell therapy is a highly personalized therapy, process of generating autologous CAR-T cells for each patient is complex and can still be problematic, particularly for heavily pre-treated patients and patients with significant leukemia burden. Here, we analyzed the feasibility and efficacy in 37 patients with refractory/relapsed (R/R) B-ALL who received CAR T-cells derived from related donors.

Patients and Methods

From April 2017 to May 2020, 37 R/R B-ALL patients with a median age of 19 years (3-61 years), were treated with second-generation CD19 CAR-T cells derived from donors. The data was aggregated from three clinical trials ( NCT03173417; NCT02546739; and ChiCTR-ONC-17012829). Of the 37 patients, 28 were relapsed following allogenic hematopoietic stem cell transplant (allo-HSCT) and whose lymphocytes were collected from their transplant donors (3 HLA matched sibling and 25 haploidentical). For the remaining 9 patients without prior transplant, the lymphocytes were collected from HLA identical sibling donors (n=5) or haploidentical donors (n=4) because CAR-T cells manufacture from patient samples either failed (n=5) or blasts in peripheral blood were too high (>40%) to collect quality T-cells. The median CAR-T cell dose infused was 3×105/kg (1-30×105/kg).


For the 28 patients who relapsed after prior allo-HSCT, 27 (96.4%) achieved CR within 30 days post CAR T-cell infusion, of which 25 (89.3%) were minimal residual disease (MRD) negative. Within one month following CAR T-cell therapy, graft-versus-host disease (GVHD) occurred in 3 patients including 1 with rash and 2 with diarrhea. A total of 19 of the 28 (67.9%) patients had cytokine release syndrome (CRS), including two patients (7.1%) with Grade 3-4 CRS. Four patients had CAR T-cell related neurotoxicity including 3 with Grade 3-4 events. With a medium follow up of 103 days (1-669days), the median overall survival (OS) was 169 days (1-668 days), and the median leukemia-free survival (LFS) was 158 days (1-438 days). After CAR T-cell therapy, 15 patients bridged into a second allo-HSCT and one of 15 patients (6.7%) relapsed following transplant, and two died from infection. There were 11 patients that did not receive a second transplantation, of which three patients (27.3%) relapsed, and four parents died (one due to relapse, one from arrhythmia and two from GVHD/infection). Two patients were lost to follow-up.

The remaining nine patients had no prior transplantation. At the time of T-cell collection, the median bone marrow blasts were 90% (range: 18.5%-98.5%), and the median peripheral blood blasts were 10% (range: 0-70%). CR rate within 30 days post CAR-T was 44.4% (4/9 cases). Six patients developed CRS, including four with Grade 3 CRS. Only one patient had Grade 3 neurotoxicity. No GVHD occurred following CAR T-cell therapy. Among the nine patients, five were treated with CAR T-cells derived from HLA-identical sibling donors and three of those five patients achieved CR. One patient who achieved a CR died from disseminated intravascular coagulation (DIC) on day 16. Two patients who achieved a CR bridged into allo-HSCT, including one patient who relapsed and died. One of two patients who did not response to CAR T-cell therapy died from leukemia. Four of the nine patients were treated with CAR T-cells derived from haploidentical related donors. One of the four cases achieved a CR but died from infection on day 90. The other three patients who had no response to CAR T-cell therapy died from disease progression within 3 months (7-90 days). Altogether, seven of the nine patients died with a median time of 19 days (7-505 days).


We find that manufacturing CD19+ CAR-T cells derived from donors is feasible. For patients who relapse following allo-HSCT, the transplant donor derived CAR-T cells are safe and effective with a CR rate as high as 96.4%. If a patient did not have GVHD prior to CAR T-cell therapy, the incidence of GVHD following CAR T-cell was low. Among patients without a history of transplantation, an inability to collect autologous lymphocytes signaled that the patient's condition had already reached a very advanced stage. However, CAR T-cells derived from HLA identical siblings can still be considered in our experience, no GVHD occurred in these patients. But the efficacy of CAR T-cells from haploidentical donors was very poor.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract